Novel Schizosaccharomyces pombe N-linked GalMan9GlcNAc isomers: role of the Golgi GMA12 galactosyltransferase in core glycan galactosylation.

Schizosaccharomyces pombe synthesizes very large N-linked galactomannans, which are elongated from the Man9GlcNAc2 core that remains after the trimming of three Glc residues from the Glc3Man9GlcNAc2 originally transferred from dolichyl pyrophosphate to nascent proteins in the endoplasmic reticulum. Prior to elongation of the galactomannan outer chain, the Man9GlcNAc2 core is modified into a family of Hex10-15GlcNAc2 structures by the addition of both Gal and Man residues (Ziegler et al. (1994) J. Biol. Chem., 269, 12527-12535). To understand the pathway of Man9GlcNAc2 modification, the Hex10GlcNAc-sized pool was isolated by Bio-Gel P-4 gel filtration from the endo H-released N-glycans of S.pombe glycoproteins. This pool yielded four major fractions, a, b, c, and g, on preparative high pH, anion exchange chromatography, that represented 10, 29, 46, and 13% of the total Hex10GlcNAc present, respectively. Structures of the glycan isomers present in each fraction were determined by one- and two-dimensional 1H NMR spectroscopy techniques. Fraction a is principally (approximately 93%) a Man10GlcNAc with a new alpha1,2-linked Man cap on the upper-arm of Man9GlcNAc. Fraction b contained two isomers of GalMan9GlcNAc in which an alpha1,2-linked terminal Gal had been added either to the upper (b1, 30%) or middle-arm (b2, 70%) of Man9GlcNAc. The gma12 - alpha1,2-galactosyltransferase-negative S. pombe strain (Chappell et al. (1994) Mol. Biol. Cell., 5, 519-528) did not make fraction b implying that the gma12p galactosyltransferase is responsible for synthesis of both isomers b1 and b2. Isomer c is Man10GlcNAc in which a new branching alpha1, 6-linked Man had been added to the lower-arm alpha1,3-linked core residue as found earlier in Saccharomyces cerevisiae and Pichia pastoris. Fraction g had less than molar stoichiometry of both Gal and Glc. The major isomer (g1, 85%) is the Man9GlcNAc core with an alpha1,3-linked branching Gal on the penultimate 2-O-substituted Man of the lower arm. This residue is also found on a novel O-linked oligosaccharide recently described in S.pombe; Manalpha1,2(Galalpha1, 3)Manalpha1,2Mannitol (Gemmill and Trimble (1999) Glycobiology, 9, 507-515). The second isomer (g2, 15%) is the partially processed Glc2Man9GlcNAc intermediate. Defining these Hex10GlcNAc structures provides a starting point for understanding the enzymology of N-linked galactomannan core heterogeneity seen on S.pombe glycoproteins.

Glycoprotein synthesis in yeast. Early events in N-linked oligosaccharide processing in Schizosaccharomyces pombe.

Oligosaccharide-lipid precursors and glycoprotein N-linked oligosaccharides isolated from the fission yeast, Schizosaccharomyces pombe, were compared with those from the budding yeast, Saccharomyces cerevisiae. Bio-Gel P-4 chromatography of oligosaccharide intermediates showed that Glc3Man9GlcNAc2-PP-dol synthesis, transfer of glycan to protein, and glucose removal to yield Man9GlcNAc2 proceeded in S. pombe as in S. cerevisiae. Two series of oligosaccharides were released from S. pombe glycoproteins by endo-beta-N-acetylglucosaminidase H; large "mannan-like" structures and smaller precursor or "core-filling" species. Unexpectedly, the smallest S. pombe N-linked glycan was Man9GlcNAc, confirmed by 500 MHz 1H NMR spectroscopy to be the lipid-linked isomer. No endoplasmic reticulum Man9-alpha 1,2-mannosidase activity was detected in S. pombe, thus identifying Man9GlcNAc as the minimum precursor for oligosaccharide elongation in contrast to the Man8GlcNAc2 intermediate identified in S. cerevisiae (Byrd, J. C., Tarentino, A. L., Maley, F., Atkinson, P. H., and Trimble, R. B. (1982) J. Biol. Chem. 257, 14657-14666). S. pombe Hex10GlcNAc was at least four isomers by high pH anion-exchange chromatography with pulsed amperometric detection. Compositional analyses identified two of the major species as GalMan9GlcNAc and GlcMan9GlcNAc, the latter of which suggests that glycan trimming may be attenuated in the S. pombe endoplasmic reticulum. Hex13GlcNAc from S. pombe was homogeneous by mass spectrometry but yielded 12 species by high pH anion-exchange chromatography. Compositional analyses, alpha-galactosidase digestion, and lectin affinity chromatography on Griffonia simplicifolia lectin I-agarose indicated these to be a family of GalxMan13-xGlcNAc isomers (X = 1-4 residues). The absence of Man9GlcNAc2 to Man8GlcNAc2 trimming in S. pombe and elongation of the lipid precursor of Man9GlcNAc with both Man and Gal to form "galactomannans" provides a novel system for N-linked glycoprotein processing studies.

Glycoprotein biosynthesis in yeast: purification and characterization of the endoplasmic reticulum Man9 processing alpha-mannosidase.

Saccharomyces cerevisiae Man9-alpha-mannosidase, responsible for trimming Man9GlcNAc2 in the endoplasmic reticulum to Man8GlcNAc2, the substrate for oligosaccharide elongation, has been purified to homogeneity from stabilized microsomal membranes without employing autolytic digestion. The activity was solubilized by the zwitterionic detergent, 3-[(3-cholamidopropyl)dimethyl ammonio]-1-propanesulphonate (CHAPS), whose presence was necessary for maximal activity. Purification included Q-Sepharose ion-exchange chromatography, preparative isoelectric focusing and HPLC gel filtration on TSK 3000 matrix. Overall purification from post-nuclear supernatants was estimated to be 110,000-fold with a 50% recovery of activity. The purified enzyme hydrolysed Man9GlcNAc1,2 from thyroglobulin or oligosaccharide-lipid, but not invertase Man9GlcNAc, Man1 alpha 2Man1 alpha OCH3 or p-nitrophenyl-alpha-D-mannopyranoside. Conversion of thyroglobulin Man9GlcNAc to Man8GlcNAc was linear with time and enzyme concentration, with an apparent Km of 0.2 mM and a specific activity of 220 IU/mg. Glc3Man9GlcNAc2 from oligosaccharide-lipid was as good a substrate as Man9GlcNAc, but the lipid-linked Man7GlcNAc2 isomer was hydrolysed at only 10% of this rate. Hydrolysis of defined isomers of IgM and bovine thyroglobulin Man6,7,8GlcNAc indicated that, for maximal alpha 1,2-mannosidase activity, only the alpha 1,2-linked terminal mannoses on the alpha 3 branch of the Man9GlcNAc precursor were dispensable. Isomers lacking the terminal alpha 1,2-linked mannose on the alpha 6 branch were hydrolysed at only approximately 10% of the maximal rate. The enzyme exhibited a pI of 5.3 and a pH optimum at 6.5. Sodium dodecyl sulphate-polyacrylamide gel electrophoresis in the absence of reducing agents gave a single sharp band at 66 kDa, while in the presence of beta-mercaptoethanol equimolar amounts of two peptides, one of 44 kDa and one of 23 kDa, were obtained. Sizing on Sephacryl SF300, Superose 12 and TSK 3000 provided a holoenzyme mol. wt of 60-68 kDa, indicating that the isolated active form of the Man9-alpha-mannosidase was composed of one each of the sulphydryl-bonded dissimilar peptides. The enzyme bound to concanavalin A (ConA)-Sepharose and was eluted with alpha-methylmannoside, indicating the presence of high-mannose oligosaccharides. The Man9-alpha-mannosidase required low levels of Ca2+, which could be removed by EGTA. Activity was restored by Ca2+ or Zn2+, but not by Mg2+ or Mn2+.

Characterization of the glycosylation sites in yeast external invertase. II. Location of the endo-beta-N-acetylglucosaminidase H-resistant sequons.

There are 14 potential Asn-X-Thr/Ser glycosylation sites, or sequons, in the yeast external invertase sequence. Of these, 13 are wholly or partially glycosylated to give an average of 9-10 oligosaccharides/subunit (Reddy, V. A., Johnson, R. S., Biemann, K., Williams, R. S., Ziegler, F. D., Trimble, R. B., and Maley, F. (1988) J. Biol. Chem. 263, 6978-6985). On digestion of native holoenzyme by endo-beta-N-acetylglucosaminidase H (Endo H) an average of about seven oligosaccharides per subunit are released without affecting enzyme activity (Trimble, R. B., and Maley, F. (1977) J. Biol. Chem. 252, 4409-4412). To determine whether the remaining Endo H-resistant chains were restricted to a limited number of unique sequons or were randomly distributed on all 13, Endo H-treated native invertase was digested with either thermolysin or trypsin and the resultant glycopeptides isolated by reversed-phase high pressure liquid chromatography and gel filtration. It was found that the oligosaccharides attached to Asn92, Asn247, and Asn350 were partially resistant to Endo H, while those at Asn45 and Asn337) were completely resistant. Bio-Gel P-4 analysis revealed the Endo H-resistant oligosaccharides at Asn45, Asn92, Asn247, and Asn337 to be Man8-14GlcNAc, while the minor residual carbohydrate at Asn350 was Man greater than 50GlcNAc. The Endo H-susceptible oligosaccharides at Asn4, Asn146, and Asn256 were Man greater than 50GlcNAc while all other glycosylation sites contained Man8-14GlcNAc. Based on a hydropathic analysis of invertase, the two most Endo H-resistant oligosaccharides at Asn45 and Asn337 were located in the more hydrophobic regions of the molecule. These may form part of the folded protein structure or interacting subunit surfaces, thus restricting their accessibility to Endo H.

Characterization of the glycosylation sites in yeast external invertase. I. N-linked oligosaccharide content of the individual sequons.

External invertase is the product of the SUC2 gene of Saccharomyces cerevisiae. The deduced sequence of this enzyme (Taussig, R., and Carlson, M. (1983) Nucleic Acid Res. 11, 1943-1954) reveals it to contain 14 potential N-linked glycosylation sites, or sequons, although only 9-10 appear to be glycosylated (Trimble, R. B., and Maley, F. (1977) J. Biol. Chem. 252, 4409-4412). To determine the location of the glycosylated sequons, external invertase was deglycosylated with endo-beta-acetylglucosaminidase H and its component peptides analyzed by both fast atom bombardment mass spectrometry (FABMS) and classical peptide isolation procedures. By use of the former technique most of the glucosamine-containing sequons could be located and by the latter sufficient amounts of small glucosamine-containing peptides were isolated to enable their quantitation. From the combined FABMS and glucosamine analyses, it was established that eight of the sequons in a subunit of invertase are either completely or almost completely glycosylated, while five others are glycosylated to the extent of about 50% or less. In the case of two overlapping sequons (4 and 5), which include Asn92-Asn93-Thr-Ser, only the first Asn was glycosylated. Thus, all but one of the sequons of external invertase are glycosylated to some extent, giving an appearance of only 9-10 N-linked oligosaccharides/subunit. The sequence identity of both external and internal invertase was verified by FABMS and by peptide sequence analysis. In only one site was an amino acid found to differ from that deduced from the DNA sequence of the SUC2 gene. This occurred at position 390 where a proline was found in place of alanine, which could result from a single base change in the triplet specifying the latter amino acid.

Amplified expression of streptomyces endo-beta-N-acetylglucosaminidase H in Escherichia coli and characterization of the enzyme product.

The endo-beta-N-acetylglucosaminidase H (Endo H) gene from Streptomyces plicatus has been cloned into the Escherichia coli plasmid pKC30 (Shimatake, H., and Rosenberg, M. (1981) Nature 272, 128-132), thus placing expression of this gene under control of the strong lambda promoter pL. The construction, pKCE3, which includes a properly positioned E. coli ribosome binding site from the lac operon (Robbins, P.W., Trimble, R. B., Wirth, D.F., Hering, C., Maley, F. Maley, G. F., Das, R., Gibson, B.W., and Biemann, K. (1984) J. Biol. Chem. 259, 7577-7583), was used to transform an E. coli strain lysogenic for a lambda prophage containing a temperature-sensitive repressor. By shifting cultures of pKCE3 lysogens to 42 degrees C, the production of Endo H commenced and was linear for about 1 h. Enzyme yields were amplified 150-fold above those obtained from comparable cultures of S. plicatus and represented 3 to 4% of total cellular protein, which enabled purification of Endo H to homogeneity by a rapid fourstep procedure. Although most of the cloned Endo H was secreted into the periplasmic space by E. coli, its 4 kDa leader sequence peptide (Robbins et al. (1984] was only partially removed during processing. As a result the purified pKCE3 Endo H was a heterogeneous population of molecules with an average molecular mass of 31 kDa compared to the 28.9 kDa fully processed product normally secreted by S. plicatus. Despite the residual approximately 2 kDa of leader sequence on the cloned pKCE3 product, there were no detectable differences in either the substrate specificity or the stability characteristics of the enzyme purified from E. coli or from S. plicatus. Of particular value for studies on glycoproteins was the finding that the genetically engineered Endo H was completely free of proteolytic contaminants.

Reappraisal of thromboplastin.

Proficiency testing surveys in the state of New York indicate that despite increased sophistication in instrumentation, there has been no real improvement in interlaboratory reproducibility in prothrombin-time determinations over the last 10 years. This lack of improvement most pronounced in the therapeutic range of 20 to 30 sec. One reason may be that between types produced by the same manufacturer. While it has been possible in several laboratories to synthesize experimentally a thromboplastin with known content of lipid and active protein, no efforts have been made to make such a product by the same manufactures. While it has been made to make such a product commercially available. Blood levels of of warfarin were measured but cannot be reliably used to monitor anticoagulation. In a preliminary study, factor Xa activity was measured using chromogenic substrate S2222. Factor Xa activity gave a positive correlation with prothrombin times of patients receiving warfarin therapy. Chromogenic substrate factor assays may represent a future method of choice for controlling anticoagulant therapy.

An analysis of electrophoretic and microcolumn methods for the separation of hemoglobins A and A2.

Parameters of three techniques for quantitating hemoglobin A2 were studied in order to identify problems affecting repeatability and then to compare intertechnique results under selected conditions. Satisfactory repeatability with cellulose acetate electrophoresis and scanning densitometry required an applicator that delivers a constant volume of sample. For cellulose acetate electrophoresis/elution and chromatographic assays a sophisticated absorption spectrophotometer and pH meter are necessary. Even with the most carefully chosen conditions significant intertechnique variation occurs. Although the colums are the most repeatable, trailing (a problem usually associated with hemoglobin electrophoresis) has also been demonstrated with column chromatography. Isoelectric focusing demonstrated the copresence of hemoglobin A and hemoglobin A2 in all trail fractions between the two major peaks and in some fractions of each peak. Standards of low protein concentration could be prepared from column fractions identified by isoelectric focusing as containing only hemoglobin A or hemoglobin A2. Such standards would be useful for assessing the accuracy of hemoglobin A2 quantitation.

Population screening for beta-thalassemia minor. Report of cooperative trials based on two approaches.

In a cooperative intrastate program based upon experience with sickle-cell anemia screening, the authors explored the feasibility of applying hemoglobin electrophoresis for detection of beta-thalassemia gene carriers. Initially, blood samples collected in capillary tubes were analyzed by cellulose acetate electrophoresis with densitometric quantitation of hemoglobin A2 (Hb A2), followed by selective spectrophotometric quantitation. This approach proved insufficiently specific or reproducible. Follow-up hematologic and family studies of presumptive beta-thalassemia gene carriers indicated that coordinate measurement of erythrocytic indices and Hb A2 values would have discriminated a subpopulation with a high incidence of beta-thalassemia trait more specifically. This approach was tested prospectively by the use of 731 venous blood samples collected in a county with a large population of Mediterranean ancestry. Of 31 individuals (4.2%) with presumptive thalassemia trait, 13 returned for a repeat testing, and the initial results for 11 were confirmed. These findings lend support to an empirical screening sequence suggested by Pearson (erythrocytic indices followed by Hb A2 quantitation), but they also indicate that a significant subpopulation of beta-thalassemia gene carriers with limited phenotypic expression may elude detection in any single-pass approach.

Endogenous cyclic adenosine monophosphate in tissues of rabbits fed an atherogenic diet.

Rabbits fed a high cholesterol diet experienced a significant increase in plasma adenosine 3,5-monophosphate (cyclic AMP), which was simultaneous with the increase in plasma cholesterol. The content of cyclic AMP in atherosclerotic lesion areas of rabbit aortic intima-media was significantly higher (0.24 picomole per microgram of DNA) than that in adjacent nonlesion areas or in aortic intima-dedia from control animals (0.09 picomole per microgram of DNA). The cyclic AMP content of heart, liver, skeletal muscle, and diaphragm showed no significant elevation in animals fed cholesterol.